In TER,
as long as the coordinates of the nodes is kept as integers, no floating
382
Chapter 15 Information Forwarding and Tra?±c Engineering
point calculations are needed. Floating point processors are expensive to
implement in hardware and are usually implemented as a set of instructions
using the integer arithmetic logic unit. In the case where floating point
is done as an algorithm, it takes many more instructions/cpu time to
compute than integer calculation.
??? With TBF, entire sections of the curve may be skipped if it has sections
which backtrack along previous sections of the curve. TER uses discrete
indexes for each point and does not encounter this problem.
??? Finally, at the cost of complexity, TBF??™s success rate is the best found
using only single-hop neighbors and path information. TER may yield a
sub-optimal solution as some of the nodes may not lie in the search window
(this is addressed by varying the size of the window).
5 Conclusions and Research Directions
In this chapter we presented a host of routing protocols and have proposed and
suggested a new one that combines expert CAD trajectory with directional
antenna/osculating window search to arrive at a route from source to destination
nodes taking into account both the proximity to the CAD trajectory as
well as the power constraints inherent in the sensor nodes. The introduction
of the CAD design, coupled with osculating window search have a double advantage
in terms of limiting the forward path to a subset of the nodes in the
network that are in the vicinity of the CAD trajectory, as well as distributing
the routing load so that the power is not depleted when information is needed
to be sent to the destination node (the sink).
Pages:
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610